Trading The Holiday Effect In Stock Markets (Anomaly and Seasonality)

Last Updated on June 30, 2021 by Oddmund Groette

The US independence holiday is just around the corner. In 2021 the fourth of July is on a Sunday, and this means Monday 5th of July is a non-trading day. We are set for a three-day holiday. However, we are traders and most of us are more interested in anomalies and seasonalities than holidays. Is there a way to profit from holiday effects, anomalies, and seasonalities? Are there holiday effects in the stock markets? If so, which holiday is the best?

In this article, we look at seven holiday effects and anomalies in the US and how the S&P 500 performs before and around these holidays. The holiday effect seems to be significant for Independence Day, Thanksgiving, and Christmas.

What is the holiday effect?

The pre-holiday effect is a seasonal effect, we can also call it a calendar effect, which tends to give increased returns in stocks over a short period of time. These situations tend to happen around holidays, like for example the 4th of July. The days before the holiday, but often also the days right after the holiday, are assumed to show consistent positive patterns and returns.

The holiday effect is widely analyzed and well known. Despite this, we know very few traders who actually trade the holiday patterns. Is there a logical reason for this? In this article, we try to find the most profitable patterns.

To those who believe such an effect will make them rich, we have to disappoint you. The holiday effect is tiny but still abnormal in a few of them. It’s not a get-rich-quick scheme but could be a useful addition to your arsenal of trading strategies.

The holiday effect is not only happening in the US. This is a phenomenon that has been documented in multiple countries across the globe.

Fundamental reasons for the holiday effect

Why is there a holiday effect?

One reason could be lower liquidity as many go on holiday. When liquidity drops, it creates a “vacuum” that means buyers need to pay up to buy shares.

Why would they pay more for the shares? The reason could be more optimism around holiday times.

Another reason could be the lack of macro news. When there is no macro news, the stock market tends to drift upwards. Volatility normally picks up when there is bad news, rarer when there is good news.

A third reason could be because participants sell their risky assets prior to the holiday season in order to reduce risk. This, of course, could lead to selling pressure and thus lower prices ahead of the holiday season. When the sellers are done, the prices gradually drift up again. As mentioned above, this creates a “vacuum”.

How to measure the holiday effect

In this article, we look at the S&P 500 because it’s the most important index in the world. We use both the ETF with the ticker code SPY but we also use ^gspc to look at data earlier than 1993. Both are downloaded for free from Yahoo!finance.

Some traders might disagree with how we test the holiday effect anomalies. That’s fine. There are no definite answers and others might have something that works better.

As always, we are always happy if others are willing to share their insights and knowledge in the comments section.

S&P 500 returns by month since 1960:

Before we start we want to have a look at the monthly returns per month in the S&P 500. This is to make a benchmark to measure the outperformance of the holiday effect.

Why is this important?

It’s important because some months are much better than others. Is the specific holiday effect good because of the pattern in that month or vice versa?

Nevertheless, here are the monthly returns in the S&P 500 from 1960 until 2021:

  • January: 1.07%
  • February: 0.08%
  • March: 0.96%
  • April: 1.56%
  • May: 0.23%
  • June: -0.01%
  • July: 0.71%
  • August: 0.22%
  • September: -0.61%
  • October: 0.94%
  • November: 1.57%
  • December: 1.33%

Clearly, the best period is from October until the end of April. This period has practically made all the gains in the market over the last 60 years. This effect we have previously documented in this article:

The Martin Luther King Jr. Day holiday effect:

The first holiday of the new year is the Martin Luther King holiday. The holiday is always on the third Monday in January and the stock market is closed to observe the day. The earliest date for the holiday is January 15 and the latest is January 21.

Let’s test the following hypothesis:

  • We go long at the close on the first calendar day of the month that is higher than 11.
  • We exit at the calendar day 21 or more.

The equity curve looks like this in SPY:

The stock market seems to get very little help from the murder of Martin Luther King, even though January is one of the best months over time. There are 29 trades, the average gain is -0.05%, the win ratio is 57%, the profit factor is 0.9, and the max drawdown is 11%.

George Washington Day/President’s Day holiday effect:

The second holiday of the year is President’s Day, officially called Washington’s Birthday, and is always on the third Monday in February. As with Martin Luther King Day, the February holiday falls between the calendar day 15 to 21.

We test the holiday effect in February this way:

  • We go long at the close on the first calendar day of the month that is higher than 11.
  • We exit at the calendar day 21 or more.

On the ETF with the ticker code SPY we get this equity curve:

The holiday effect is absent in February. There are 29 trades, the average gain is -0.24%, the win ratio is 52%, the profit factor is 0.77, and the max drawdown is 9%.

The Easter holiday effect:

The next holiday after Washington Day in February, is Good Friday during Easter. We chose to skip this holiday because it varies in both month and calendar day, and hence very hard to code.

Thus, we skip it.

The Memorial Day holiday effect:

Memorial Day, the day to honor the men and women who died while serving in the US military, is always on the last Monday of May.

Let’s test the following hypothesis:

  • We go long at the close the Wednesday prior to Memorial Day.
  • We exit at the close on the first trading day of June.

In SPY the Amibroker equity curve looks like this:

The chart clearly indicates the effect is small to non-existent. There are 29 trades, the average gain is 0.18%, the win ratio is 59%, the profit factor is 1.17, and the max drawdown is 5%.

The 4th of July holiday effect (Independence Day effect):

The 4th of July is a public holiday and the markets are closed. If the 4th of July is on a weekend, the markets are closed on the following Monday.

How does the stock market perform up to Independence day?

Let’s test the following hypothesis:

  • We buy the S&P 500 at the close of the second last trading day of June. For example, we buy at the close on the 29th of June so we are long on the open of the last trading day of June.
  • We sell six days after we bought. The effect seems to last a few days after the holidays, thus we keep the position a bit longer.

Now, before considering the strategy, keep in mind that June is a poor month for stocks. Additionally, there is a strong tendency for the markets to perform much better in the last days of the month and the first few days of the new month. Please check out this article:

The strategy returns this equity curve in SPY:

There are 28 trades, the average gain is 1.01%, the win ratio is 68%, the profit factor is 2.97, and the max drawdown is 4%. This equals a CAGR of 0.5% while being invested just 0.6% of the time.

The returns per year look like this:

Several tests with different entry and exit dates indicate the effect is pretty good.

The Labor Day holiday effect:

Labor Day is always on the first Monday of September. What has the performance been around this day?

Let’s the following strategy:

  • We buy the S&P 500 at the close of the second last trading day of August. For example, we buy at the close on the 30th of August so we are long on the open of the last trading day of August.
  • We sell at the close on the first Tuesday of September (after the first Monday).

The equity curve in Amibroker looks like this (100 000 compounded since the start):

There are 28 trades, the average gain is 0.11%, the win ratio is 57%, the profit factor is 1.14, and the max drawdown is 5%.

The result is pretty poor.

We tried both buying earlier and later but the results seem pretty random.

Keep in mind that September is historically the worst month of the year:

The Thanksgiving holiday effect in the S&P 500:

What about the famous turkey celebrations in November?

Thanksgiving is always on the fourth Thursday in November and began as a day of blessing for the harvest in the preceding year. The day is a non-trading day and the next day, Friday, is only half a trading day where the markets close at 1 pm Eastern Time. The Friday is called Black Friday because of its importance for the retail sector.

We test the holiday effect of Thanksgiving this way:

  • We go long at the close the Monday prior to Thanksgiving.
  • We exit at the close on the first trading day of December.

By testing on SPY Amibroker returns this equity curve:

There are 28 trades, the average gain is 0.82%, the win ratio is 65%, the profit factor is 2.72, and the max drawdown is 5%. This equals a CAGR of 0.4% while being invested just 0.6% of the time.

The Thanksgiving effect has been pretty consistent since 1960 (ticker code is ^gspc – the cash index of the S&P 500):

The Thanksgiving holiday effect in the retail sector (Black Friday effect):

Because of Black Friday, which has developed into a consumer bonanza, it might be interesting to check the performance of the retail sector. We test by using the Fidelity Select Retailing fund (FSRPX). We use the same criteria as we did in the Thanksgiving holiday effect for the S&P 500.

The retailers perform slightly better than the S&P 500:

There are 35 trades, the average gain is 1.14%, the win ratio is 71%, the profit factor is 3.24, and the max drawdown is 6%.

Keep in mind, though, that November and December are strong months in the stock market.

The most wonderful time of the year: The Christmas holiday effect (Santa Claus rally)

One of the best periods of the year is just after Christmas and the first days of the next year. We have covered this in a separate article where we tested the following:

  • Go long at the close of the fourth last day of the year
  • Sell at the close on the third trading day of the new year.

 

In practice, this is the same as the turn of the month strategy.

Let’s make a twist where we go long at the close of the first trading day after the 20th of December and sell on the first trading day of the new year.

The backtest in Amibroker yields this equity curve since 1960:

There are 60 trades, the average gain is 1%, the win ratio is 67%, the profit factor is 4.8, and the max drawdown is 2.9%. This is pretty solid results!

Relevant articles about seasonalities and anomalies:

There are almost endless anomalies and seasonalities in the stock market, and also in other markets. Most of them are not tradeable, but some are. We have covered several of these:

The tests in this article were done in Amibroker. If you would like to know the code of our tests you can order them here and at the same time get the code for at least 60 other strategies we have published on this website:

Conclusions about the holiday effect anomalies:

Of the seven holiday anomalies and seasonalities we did in this article, only three seem to be tradeable: Independence day, Thanksgiving, and Christmas.

However, all the strategies can be changed and tweaked. We have some ideas to improve the three best holiday seasonalities, something we will come back to later.

 

Disclosure: We are not financial advisors. Please do your own due diligence and investment research or consult a financial professional. All articles are our opinions – they are not suggestions to buy or sell any securities.